Comment 4 for bug 641899

Hi everyone! I am entirely agreed that the startup time of s-c is something that needs improving. This is something that I am dedicated to 'fixing' during the natty cycle. Some of you may know that I worked on this a bit during the summer; In an experimental branch I managed to reduce warm startup time from about 3s to 1s and cold startups from 14s to 8s. The branch is unsuitable for merging, but it allowed me to get a feeling for where the slowness lies and what to do about it. Since then much of the code has changed (eg the startup speed of trunk has increased due to new features), so I will need to do the work again - although I will make similar changes to what I made back then (and also a few other crucial changes which I have come up with since then). I think that 1.5s and 10s startup times are realistically achievable without creating a 'minimal' interface. 1s and 5s startup times are the ultimate goal which should be achievable with further tweaks to the existing model. This is just a quick message to let you know that the issue of startup speed isn't being ignored - it will probably be December before I have time to work on this though. Feel free to ping me at UDS about any of this. Cheers.